電漿處理對二氧化鉿/矽介面層抗輻射能力之影響

Abstract

極紫外光微影是最可能成為下個世代微影技術的選擇，但其光源屬於高能量輻射，若要應用於工業製程上，尚須考慮游離輻射對於電子元件所造成的影響。前人已利用金屬閘極/高介電常數介電質金氧半電容器研究不同材料、介面厚度、氧化層厚度和退火溫度之影響，但仍尚未透徹了解高介電質材料在極紫外光照射下的損傷。因而本論文研究著重於介面品質差異性之抗輻射能力，利用氨氣與氮氣電漿對介面層進行品改善。另外，我們也利用先進製程N型電晶體，研究經由高溫退火修復元件輻射損傷至照射前電性，是否降低元件之抗輻射能力，以及極紫外光照射對元件的正偏壓溫度不穩定性的可靠度分析。 先前研究成果已證實，游離輻射會在金氧半元件的閘極介電層中產生正電荷、邊緣缺陷和介面缺陷，從元件的電容-電壓曲線可看到位移、遲滯電壓增加和曲線變形等現象。經由電漿處理後發現，氮氣電漿試片經極紫外光照射後，平帶電壓增加量相對於對照組變化較少，且遲滯電壓與之相差無幾，介面缺陷變化是三者中最少，較適合採用於製程中。另外，氨氣電漿處理會因為大量氫離子摻入，使元件電性衰退，造成嚴重的電容-電壓曲線位移與變形，因此並不適合採用。且在5奈米氧化層試片，氨氣電漿試片之平帶電壓與介面能態增加量較多，但遲滯電壓所有條件差異性不大。接著採用10 keV之X-光，由於衰退長度與極紫外光相差非常大，以5奈米二氧化鉿為例，換算出X-光的吸收率與極紫外光相差約110倍，因此各照射條件下，電性變化並不明顯。由上敘述可推論，三種條件為氮氣電漿試片抗輻射能力最佳，而氨氣電漿試片最差，當氧化層調降為5奈米，雖然輻射影響大幅下降，氨氣電漿試片仍為最不適合採用的電漿處理條件。 最後，我們探討退火對抗輻射力之影響，電晶體接受不同能量輻射照射，之後經由高溫退火並再次照射輻射，證實其高溫退火可修復輻射損傷，但不會影響抗輻射能力。然後我們利用臨界電壓對正偏壓溫度的不穩定性，於較弱之電場下測試輻射對元件可靠度之影響，得知結果仍無法明確肯定輻射對可靠度之影響，但證實於測試中，電場影響遠大於輻射影響。

Extreme ultraviolet lithography (EUVL) is the most likely selection as the next generation lithography technology in the future. The radiation damage effect during exposing process should be considered because the energy of EUV is higher than chemical bonding energy of most dielectrics. The effects of oxide thickness, interfacial layer thickness, gate oxide material, and annealing temperature on the radition hardness of high-k/metal gate MIS capacitors have been studied previously. In this thesis, the radiation hardness of the HfO2/Si structure with plasma-treated interfacial layer is investigated. Moreover, we prepare the state-of-the-art n-channel Metal-Oxide-Semiconductor-Field-Effect-Transistor (nMOSFET) to study the effect of post-radiation annealing. In addition, we use positive bias temperature instability (PBTI) test to evaluate the radiation effect on the reliability property of nMOSFET. Previous study reported that ionizing radiation would induce positive oxide charge, interface traps, and border traps in the gate oxide of MIS capacitor. These defects cause capacitance-voltage (C-V) curve shift, C-V curve distortion, and increment of hysteresis. After EUV irradiation, it is observed that N2 plasma treatment increases the flatband voltage slightly. Furthermore, the increment of hysteresis is similar to that of the control sample and the interface traps increase is less than other samples. The experiment results imply that the N2 plasma treatment would improve the interface property by nitrogen incorporation. Besides, it is well suited for industrial processes. On the other side, the electrical characteristics of the NH3 plasma treated sample degrade after ionizing irradiation due to hydrogen incorporation. Evident shift and distortion of the C-V curve is observed. These results indicate that the NH3 plasma treatment is not suitable to be used in industrial processes. As the radiation source changes to 10 keV X-ray, the effect of radiation damage on the C-V curves is unobvious because the percentage of photons which are absorbed by HfO2 layer for EUV is 110 times larger than that for the 10 keV X-ray. In conclusion, N2 plasma treatment is the best choice. The worst case is the NH3 plasma treatment. As the thickness of HfO2 decreases to 5 nm, the degradation of flatband voltage and hysteresis are reduced apparently. Similarly, the N2 plasma sample still has the least flatband voltage shift and interface state generation. The worst case is NH3 plasma sample, too. Finally, we discuss the effect of annealing on radiation hardness. The state-of-the-art nMOSFET is irradiated by different radiation source. Next, it is annealed at 400 ℃. After annealing, it is exposed again to the same dose irradiation. Experimental results indicate that high temperature annealing could recover the radiation damages. Moreover, the annealing wouldn’t change the device radiation hardness. Afterward we evaluate the PBTI at weak electric field. We observe that no matter what radiation source is used, the distribution of threshold voltage degradation is nearly the same both at room temperature and 150 ℃. It indicates that the reliability property of the state-of-the-art nMOSFET wouldn’t be affect by ionizing radiation.